Scholarlink

This paper proposes a 2D Logistic-Sine-Hénon Chaotic System (2D-LSHCS) to enhance image data security in wireless networks. The system demonstrates superior chaotic performance, with a Lyapunov exponent of 47.4795, sample entropy of 1.9376, and a 0–1 test score of 0.986. The encryption scheme incorporates dynamic key updates and a novel embedding method that combines Discrete Cosine Transform (DCT) with phase adjustment, increasing both security and data hiding capacity. Optimization is achieved through a reinforcement learning approach based on Bayesian theory. Experimental results under 3BPP embedding show an NPCR of 99.61%, UACI of 33.45%, PSNR of 47.50, and SSIM of 0.9973, confirming a strong balance between high security, image quality, and steganographic capacity.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Adaptive Reinforcement Learning-Driven Image Hiding Scheme with Chaotic Encryption

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Tingting Shao,
  • Shuhao Zhang,
  • Yifan Quan,
  • Haojin Jiao,
  • Man Zhou

摘要

This paper proposes a 2D Logistic-Sine-Hénon Chaotic System (2D-LSHCS) to enhance image data security in wireless networks. The system demonstrates superior chaotic performance, with a Lyapunov exponent of 47.4795, sample entropy of 1.9376, and a 0–1 test score of 0.986. The encryption scheme incorporates dynamic key updates and a novel embedding method that combines Discrete Cosine Transform (DCT) with phase adjustment, increasing both security and data hiding capacity. Optimization is achieved through a reinforcement learning approach based on Bayesian theory. Experimental results under 3BPP embedding show an NPCR of 99.61%, UACI of 33.45%, PSNR of 47.50, and SSIM of 0.9973, confirming a strong balance between high security, image quality, and steganographic capacity.